5292
J . Phys. Chem. 1988, 92, 5292-5297
This is in contrast with the observations in the monomeric methoxybenzene positive ion radicals, which in some cases have been found to be powerful oxidants, with no reducing proper tie^.'^,^^ The difference is attributed to the structure of the polymeric radical ion, which enables redox reactions involving the removal of the tertiary hydrogen from the C H group adjacent to the benzene ring.
Acknowledgment. This research was supported by the Israel US BNSF and by the Balfour Foundation. We are indebted to E. Gilead for assistance with the maintenance of the pulse radiolysis setup. Registry No. PSS, 50851-57-5; HO', 3352-57-6; Na2S04,7757-82.6; MgSO4, 7487-88-9; 0 2 , 7782-44-7; Cu(CIOJ,, 13770-18-8.
Thermolysis of a Polymeric Endoperoxide: The Yield of Singlet Oxygen Released into the Gas Phase Allen Twarowski* Rockwell International Science Center, Thousand Oaks, California 91 360
and Phan Dao Air Force Geophysics LaboratorylLID, Hanscom AFB, Massachusetts 01 731 (Received: December 21, 1987)
The fraction of oxygen that is electronically excited upon release from thin films of 1,4-dimethyl-2-poly(vinylnaphthalene 1,4-endoperoxide) was measured by reaction with 2,5-dimethylfuran, DMF. The amount of DMF that reacted with singlet oxygen was assayed by mass spectrometry and the dependence of the singlet oxygen yield on film thickness was found to be in substantial agreement with previous experimental results. Comparison of these measurements with model calculations is reported.
Introduction Thermolysis of the endoperoxides of many polycyclic aromatic hydrocarbons has been shown to result in efficient conversion of thermal energy to electronic excitation of the product oxygen molecule.' Recently, solid films of 1,4-dimethy1-2-poly(vinylnaphthalene 1,4-endoperoxide), 2PVNE, were reported to quantitatively release oxygen into the gas phase upon heating. The fraction of released oxygen which was electronically excited to the IAe was found to vary markedly with 2PVNE film thickness in agreement with a simple model of molecular diffusion and quenching in the polymer solid. The O2 'Ag fraction was determined from luminescence measurements at 1270 nm using a liquid nitrogen cooled germanium photodiode.2 There are difficulties with the determination of singlet oxygen yields from photometric data. First, the photometric signal is proportional to the number of luminescing species present at any one time rather than the integral amount of species produced. This problem was circumvented in our previous work2 by releasing the singlet oxygen quickly into a 10 Torr buffer gas which served to confine the excited oxygen to the viewing region of the detector for a period of time comparable to that required for thermolysis of the 2PVNE and release of O2from the polymer film. Under these conditions the peak germanium detector signal is proportional to the total amount of O2 IAg released into the gas phase. Another difficulty with photometric assay of O2 derives from the large number of physical constants required to convert the germanium detector signal to an absolute measure of luminescing species population. Uncertainties associated with the special distribution of O2 lAg in the sample chamber, the effective f number of the optical setup, the responsivity of the detector, and the radiative rate constant for the 02(1A,-3Z,) transition all contribute to the uncertainty of the calculated singlet oxygen yield. A more direct measure of the O2 'A, yield can be obtained by preferentially reacting the singlet oxygen molecule with another chemical species and assaying the reactants or products. This
'$
(1) Turro, N. J.;
Chow, M. F.; Rigaudy, J. J . Am. Chem. SOC.1981, 103,
7218. (2) Twarowski, A. J.; God, L.; Busch, G. E. J . Phys. Chem. 1988,92, 396.
0022-3654/88/2092-5292$01.5O/O
method has been used to measure singlet oxygen density in the gas flow from a microwave discharge source. Gleason et aL3 measured the rate constants for the reaction of O2 lAg with tetramethylethylene (TME) and with 2,s-dimethylfuran (DMF) using a gas-liquid chromatograph to separate reactants from products and quantify the amount of material reacted. Their findings suggested that DMF physically quenched O2 'A, at about the same rate with which it reacted, whereas TME, though having a reaction rate with O2 lA, which was about 4 times slower than DMF, reacted quantitatively. The rate constants reported by these workers for the reaction of O2 'Ag with DMF and TME are 3.7 X lo8 and 1 X lo8 cm3 mol-' s-I, respectively. Huie and Herron4" measured the rate constants for the same reactions of singlet oxygen with D M F and TME using mass spectrometry to detect the reactants and products. D M F was found to react quantitatively with O2 lA,, in contrast with results reported by Gleason et al.3 The rate constants which Huie and for DMF Herron repod are 1.5 X 1Oloand 7.7 X IOs cm3 mol-I and TME, respectively. These rate constants are significantly larger than earlier reported values and the differences have been attributed to the presence of atomic oxygen in the gas flow in earlier studies. The effects of atomic oxygen were minimized in the studies reported by Huie and Herron by addition of NO2 to the singlet oxygen flow.5 In this paper we report the fractional yield of O2 lABreleased from 2PVNE films determined by reaction of the excited oxygen with TME and DMF. Using mass spectrometry to assay the quantity of D M F reacted, we found the singlet oxygen yield to be in substantial agreement with previous photometric determinations. Experimental Section The preparation of ZPVNE, its deposition as thin films on resistive glass plates, and the heating of the endoperoxide films (3) Gleason, W. S.; Broadbent, A. D.; Whittle, E.; Pitts, Jr., J. N. J . A m . Chem. SOC.1970, 92, 2068. (4) Herron, J. T.; Huie, R. E. Ann. N . Y . Acad. Sci. 1970, 171, 229. (5) Huie, R. E.; Herron, J. T. In?. J . Chem. Kine?. 1973, 5, 197.
0 1988 American Chemical Society
The Journal ofPhysical Chemistry. Vol. 92, No. 18. 1988 5293
Thermolysis of a Polymeric Endoperoxide
I
IAb",
"Yrn
r
Con,sr,r
Resllue NU.# o"#de Coaling Glass S Y I 1 I D I P
Figure 1. The heated substrate on which ZPVNE is deposited.
0 Rl"0 sa61
were described in detail in ref 2. For the present series of experiments 2PVNE was spin-coated directly on the electrically resistive metal oxide layer (about 30 ohms per square) of in. thick Nesatron plates (Nesatron is a glass substrate with an electrically resistive coating obtained from PPG Industries). The passage of an electrical current through the resistive Nesatron coating heated the surface of the glass plate and the adjacent endoperoxide film. The thermolysis area was limited to 3.2 cm2 by etching 1.5-in. square Nesatron plates in hydrochloric acid which left a resistive strip in the center of the plate as shown in Figure I . Aluminum strips 1 pm thick were vacuum deposited on two sides of the thermolysis area providing a high conductance current path to the electrically resistive metal oxide coating in the center of the plate. Copper rods in physical contact with the aluminum strips were connected to a Kepco ATE 350-3.5111 programmable power supply which provided the electrical power needed to heat the thermolysis area of the Nesatron plate. For one series of experiments the resistive coating of the Nesatron plate was overcoated with 0.1 pm of silicon monoxide to provide an insulating surface on which the endoperoxide was then deposited. The experimental apparatus which was used to measure the singlet oxygen yield upon thermolysis of the endoperoxide films is illustrated in Figure 2. A sample chamber with an internal volume of 14.5 cm3 was constructed from Teflon to minimize deactivation of 0,'$on the walls of the container. The Nesatron plate coated with 2PVNE was mounted to the sample chamber and the central thermolysis section was sealed to vacuum with an 0 ring. The sample chamber along with the gas-handling manifold was evacuated with a 150 L/s diffusion pump and filled with roughly 50 mTorr of reactant (DMF or TME) and 0.5 Torr of N,. The nitrogen buffer gas was added in order to slow the rate of diffusion of singlet oxygen to the walls of the sample chamber. The pressure in the manifold was accurately measured after each gas addition with a 1 Torr Baratron capacitance manometer (Type 127AA) calibrated by the manufacturer (MKS). After addition of reactant (DMF or TME) and buffer gas, the sample chamber was isolated from the gas handling manifold with a Teflon valve and the sample substrate was then heated by passing a current through the resistive metal oxide coating resulting in the thermolysis of the adjacent 2PVNE film. A 25 pm diameter pinhole mounted to a port in the sample chamber provided a slow leak of the sample gas to the ionization section of the UTI Model IOOC quadrupole mass spectrometer equipped with an electron multiplier ion sensor. The detector section of the mass spectrometer was evacuated with a Leybold-Heraeus turbopump (Turhovac 50). Tetramethylethylene (Aldrich Chemical) and 2Jdimethylfuran (Aldrich Chemical) were purified by distillation and kept refrigerated until just prior to use. The polymeric endoperoxide, 2PVNE, was prepared from the monomer and peroxidized as described in ref 2.
'Is
Figwe 2. The experimental apparatus for measuring singlet oxygen yield upon thermolysis of ZPVNE.
Results 1 . Single Film Thickness: DMF. The total amount of molecular oxygen (both ground state and electronically excited) released into the gas phase upon thermolysis of a 19 nm thick ZPVNE film was measured for a series of six samples. The film was heated to 88 OC for 5 s and the change in the ion current at m / e 32 was recorded. The mass spectrometer was separately calibrated for 0,by rwording both the ion current at m / e 32 and the sample chamber pressure as pure oxygen was metered into the sample chamber. The number of moles of 0,released into the gas phase was determined by first multiplying the mass spectrometric ion current change at m / e 32 by the calibration factor to give the partial pressure of 0,in the sample chamber and then using the ideal gas law (the gas was assumed to he thermally equilibrated with the sample chamber walls). Dividing the number of moles of 0,thus calculated by the area of the 2PVNE thermolysis region (3.2 cm2) gave (5.33 f 0.22) X IO4 mollcm' for 19 nm thick 2PVNE films. The number of moles of 0,released from the 2PVNE film per unit area can be independently determined from the following data: ( I ) the fraction of 2PVNE that has been converted to endoperoxide (this is measured by solution phase absorption spectrophotometry), (2) an average density of 1 g/cm3 (supported by data in ref 2), (3) the average monomer molecular weight of the polymer, and (4) the film thickness determined from the coating solution concentration (using a previously determined correlation of 8.4 nm of film thickness for each 1 g of 2PVNE per liter of solution).' This calculation gives an 0,yield of 6.27 X IO* mol/cm2 which is about 15% larger than the value determined by mass spectrometric assay of 0,. The fraction of 0,released in the 'Asstate upon thermolysis of 2PVNE was measured by reaction with either DMF or TME. The extent of the reaction was determined by observing the change in time of the mass spectrometer ion current when the quadrupole filter was set to pass a particular mass to charge ratio ( m / e ) associated with the 02+ ion or with a reactant or product fragment ion. From the mass spectrometric signal, the number of moles of 0,IAs released into the gas phase, n*, was calculated. The fractional yield was then obtained by dividing n* by n, the total number of moles of O2 released. The value of n is found by assuming that it is directly proportional to film thickness and using the value of 5.33 X 10-9 mol/cm2 reported above for 19 nm thick films. When DMF is used as a reactant, the amount of singlet oxygm released into the gas phase can be independently determined from the data acquired at several mass peaks, including the primary product signal at m / e 128, the primary 0,signal at m / e 32, and the primary DMF signal at m / e 96. The appearance of a product
5294
I warowsKi
The Journal of Physical Chemistry, Vol. 92, No. 18, 1988 L.V
I1
I
-
w -
5- 2.2 -
,
0
u.
5
ana u a o
..
. . .
-
1.8 1.4 -
z
z! TIME
Figure 3. The decrease in ion current at the primary DMF mass peak (96) upon release of singlet oxygen from ZPVNE: (A) before thermolysis, (B) after thermolysis, and ( C ) after sample chamber evacuation.
peak in the mass spectrum at m l e 128 upon titration of singlet oxygen by DMF has been reported by others4 but was not observed in our experiments. Changes in the ion current at m l e 32 and m l e 96 were observed, however, and these changes were used to measure the 0, 'A, yield. When D M F is absent from the sample chamber, the change in ion current a t the primary O2 mass peak ( m l e 32) upon thermolysis of ZPVNE, s0(32), is proportional to the total amount of O2 released (both electronically excited and ground state). When DMF is present in the sample chamber, the primary O2 signal is diminished in magnitude from S0(32). With an excess of DMF, the change in ion current, S(32), is proportional to the sum of the 0, released into the gas phase in the ground state and that O2 released into the gas phase in the singlet state which is quenched to the ground state without reaction with DMF. The latter was assumed to be negligible. The fractional yield of 0, 'A, released into the gas phase is equal to 1 - S(32)/S0(32). From ion current changes at m l e 32 the fractional yield was calculated to be 0.40 0.08 for 19 nm thick 2PVNE films heated to 88 OC for 5 s. The determination of this quantity can be in error if either the reactant (DMF) or the reaction product fragments to give an ion current at m l e 32 and if these contributions are not equal. The O2 'A, yield also can be determined from the change in ion current at the primary DMF mass peak upon release of singlet oxygen into the sample chamber. This method is less troubled by possible competing fragmentation channels than is the determination using the primary O2 mass peak. The ion current at m / e 96 was measured over the course of the 2PVNE film thermolysis and the fractional change was computed. The initial partial pressure of D M F was measured with a capacitance manometer and the number of moles of DMF in the sample chamber was computed using the ideal gas law. The latter when multiplied by the fractional change in m l e 96 ion current gives the number of moles of D M F consumed in the reaction with singlet oxygen. The number of moles of O2 'As released into the gas phase, n*, is taken to be equal to the DMF consumed in the reaction. The fractional yield of O2 IAg is simply the ratio of n* to the total number of moles of oxygen released. For these measurements n was determined in a separate set of experiments in which DMF was absent from the sample chamber and the change in the ion current at mle 32 was recorded (vide supra). The calculated value of the fractional O2lA, yield using this data is 0.30 f 0.02 when a 19 nm thick 2PVNE film is thermolyzed by heating to 88 OC for 5 s. This value obtained from measurements of the primary reactant mass peak compares with 0.40 0.08 which was obtained from measurements of the ion current at the primary oxygen peak ( m l e 32). The fractional yield which is calculated from the m / e 96 data will underestimate the true fraction of O2 lA, released into the gas phase if the product (presumed to be an endoperoxide of DMF) formed by reaction of singlet oxygen with DMF fragments in such a manner so as to give an ion with a mass to charge ratio of 96. The extent of this problem was assessed by reacting all the D M F in the sample chamber with singlet oxygen to give DMF-O, and measuring the change in m l e 96 ion current. Since the thermolysis of a 19 nm thick 2PVNE film results in the release of enough singlet oxygen to consume about 6 mTorr of DMF, the
*
*
I
5
o.8 0.4
i
" 0 81
83
85
87
89
91
TEMPERATURE (C)
Figure 5. The number of moles of DMF consumed upon release of singlet oxygen from ZPVNE per unit area of endoperoxide film as a function of thermolysis temperature.
addition of only 2 mTorr of DMF to the sample chamber ensured that ample O2I$ was present during the course of the experiment to react with all the DMF. Figure 3 shows the result of this experiment. The ion current ( m l e 96) drops to about 5% of its initial value as O2lA, reacts with almost all of the D M F present in the sample chamber. If the DMF-singlet-oxygen product contributed significantly to the ion current at m / e 96, a dramatic decrease in the mass spectrometric signal such as seen in Figure 3 would not be observed. Therefore, the value of the O2 'A, yield calculated from the fractional change in ion current at m l e 96 is probably not reduced from the true value by more than 5% as a consequence of fragmentation of DMF-02. The O2 'A, yield can be measured only when an excess of reactant (DMF or TME) is present in the sample chamber and under these conditions the amount of D M F which reacts with O2 lA, should be independent of the initial amount of DMF. Several experiments were performed with differing amounts of DMF (at least twice the amount required to fully react with all the O2l A, released) loaded into the sample chamber to substantiate this conjecture. Figure 4 shows the number of moles of DMF consumed upon release of O2lA, as a function of initial DMF partial pressure. The amount of DMF consumed in the reaction is not significantly affected by the initial partial pressure of DMF as long as at least a twofold excess of D M F is initially present. The amount of DMF consumed upon thermolysis of a 19 nm thick 2PVNE film was measured over a fairly restricted range of substrate temperature, 81-92 OC, to assess whether small variations in temperature which may occur among a series of experiments can result in significant changes in the measured singlet O2 lAg oxygen yield. Figure 5 shows the variation of the amount of DMF consumed as a function of substrate temperature. From 81 to 88 O C this fraction does not change. A small increase may occur when the temperature is raised to 92 "C but the variation in the experimental data is comparable to the observed increase. 2. Single Film Thickness: TME. TME has previously been used as a titrant for O2 'A,; however, because the reaction with 0, IAg is much slower than with DMF, more opportunity exists for singlet oxygen deactivation either by collisions with TME, collisions with the buffer gas, or by deactivation on the walls of the sample chamber (particularly the endoperoxide film surface). In a series of experiments TME replaced DMF as the chemical
The Journal of Physical Chemistry, Vol. 92, No. 18, 1988 5295
Thermolysis of a Polymeric Endoperoxide
000
100
200
300
400
500
,000
FILM THICKNESS (nm)
Figure 6. The fraction of oxygen released into gas phase in the 'A state as a function of ZPVNE film thickness: 0 , experimental results determined from mass spectrometric analysis of the primary DMF mass peak -, model calculation results (see text).
reactant in the 0, IAg yield determination. An increase in ion current at m / e 83 was observed upon thermolysis of 2PVNE and was attributed to the C6HII+ion resulting from the fragmentation of the addition product of 0, lAg and TME. A decrease in ion current upon reaction of TME with singlet oxygen was observed at the primary TME peak at m l e 84 and at the fragment peak at m / e 69. The yield of O2'Ag from thermolysis of 19-nm 2PVNE films heated to 88 O C for 5 s was calculated independently from the fractional decrease in ion current at m l e 84 and 69 and was found to be 0.1 1 in both cases. This compares with 0.30 measured by the same procedure (fractional decrease in reactant ion current) with D M F as the chemical titrant. 3. Variable Film Thickness. In previous experiments in which a germanium detector was used to photometrically determine the yield of 0, 'A, it was found that the fraction of singlet oxygen released into the gas phase varied markedly with film thickness. The primary motivation for the present series of experiments was to verify these measurements with a technique which could give a more accurate measure of the 0, 'Ag yield. The fractional yield of O2 IA, released into the gas phase upon thermolysis of 2PVNE at 88 O C was measured as a function of film thickness by using D M F as the chemical titrant and recording the change in ion current at m / e 96. The results of this study are plotted in Figure 6 for the range of film thickness of 0-50 nm. The error bars attached to these results are the standard deviations from the mean of at least five samples for each film thickness. The solid curve in Figure 6 was calculated from a diffusional model described in ref 2 which includes a pseudo-first-order rate constant for the quenching of O2 'As by the polymer matrix as the 0, 'Ag diffuses through the polymer film. The fraction of oxygen released from a film of thickness Tin the 'A, state is given by
(1) The ratio of the quenching rate constant, k,, to the diffusivity of 0, lA, in the polymer film, D, is an adjustable parameter of the model. Two other parameters implicit in the model as described in ref 2 are explicitly included in eq 2 above; the probability, y, that a single oxygen molecule will be quenched at the 2PVNE/Nesatron interface and the intrinsic yield, 4, of singlet oxygen upon thermolysis of the endoperoxide moiety in the polymer. In principle both y and 4 in eq 1 can range from 0 to 1; however, the asymptotic value of the measured fractional yield at zero film thickness is close to 0.5 and therefore limits the intrinsic yield to values between 0.5 and 1. The incorporation of the (1 - (y/2)) factor in eq 1 is not rigorously justified here except at the limiting values of 0 and 1 which correspond with the elastic and inelastic models discussed in ref 2. The solid curve in Figure 6 was calculated assuming that both y and 4 are unity. A good fit with the experimental data is achieved when k,/D =
60.0
120. 180. 240. FILM THICKNESS (nm)
300
Figure 7. The fraction of oxygen released into the gas phase in the 'A state as a function of ZPVNE film thickness: 0 , experimental results
determined from photometric data; -, model calculation results (see text).
f
5E m0L
L
300+
t
i
A
1 1 1 40.0
,000
10.0
50.0
20.0 30.0 FILM THICKNESS (nm)
Figure 8. Fractional yield of oxygen 'A as a function of ZPVNE film thickness. Experimental results were determined from mass spectrometric analysis of the primary DMF mass peak. *, ZPVNE/Nesatron interface; 0 , ZPVNE/SiO/Nesatron interface.
2 X lo1, cm-*. An equally good fit is obtained for 4 = 0.5 and y = 0. However, inspection of eq 1 shows that with these choices for and y,the adjustable parameter, k,/D, must be reduced by a factor of 4 to give the same dependence of the fractional yield, F, on film thickness. The results of O2 lAg yield measurements obtained by chemical reaction with DMF (measured by following the decrease in ion current at m / e 96) compare favorably with the O2 lAg yields obtained from germanium detector measurements previously reported. Figure 7 shows the germanium detector measurements of the fractional 0, IAg yield as a function of film thickness but over a larger range of film thickness (10-250 nm) than was studied with the chemical reaction method. The data are reasonably well fit by model calculation results (solid curve in Figure 7 ) using the same value of k , / D and the same choices for y and 4 as was used for the model calculation results of Figure 6. It is not possible to determine the extent of quenching of singlet oxygen at the polymer/Nesatron interface from the fitting of model calculations with the measured O2 I Ag yields as inspection of eq 1 makes clear. Equally good agreement is obtained at the extremes of the range of y (0 and 1) if 4 is allowed to vary from 0.5 to 1.0. It may be argued that the conducting metal oxide coating on Nesatron contains a wide distribution of impurity charge carrier trapping sites as well as mobile electrons which could interact with O2 lAg providing an unusually effective quenching surface. This argument favors a quenching probability and an intrinsic yield both close to unity. An experiment was performed to determine the extent to which charge carriers in the Nesatron coating act as quenchers of O2 lAs. A silicon monoxide coating of 100 nm was deposited on top of the resistive metal oxide coating of Nesatron and the endoperoxide film was spin-coated on top of this electrically insulating layer. Figure 8 shows the fractional yield of 0, 'A, as a function of 2PVNE film thickness obtained when the polymer was coated on silicon monoxide and also includes the measurements presented in Figure 6 where 2PVNE was deposited directly on the bare Nesatron surface. No significant difference between the two sets of measurements is apparent suggesting that if the Nesatron
5296
The Journal of Physical Chemistry, Vol. 92, No. 18, 1988
surface is in fact an effective quencher for O2 IA, the quenching mechanism probably does not involve charge carriers. Summary and Discussion The amount of 0, released by a 19 nm thick film of 2PVNE mol/cm2. determined by mass spectrometric assay is 5.33 X This value is about 15% below that which is calculated from film thickness, mass density, and fractional conversion of the naphthalene moieties to endoperoxide in the film. A 15% disagreement in total oxygen yield is not unreasonable considering the complexity of the experiment and the number of factors included in the calculations. Possible sources of error are the estimates for (1) the area of the 2PVNE film which is heated (this area was taken to be equal to the area of the etched Nesatron metal oxide coating), (2) the film thickness, (3) the mass density of the film, and (4) the percent conversion of naphthalene to endoperoxide. It should be noted that the 0, 'A, fractional yield measurements reported here (as distinguished from the measurements of the total amount of 0, released) are not subject to systematic error arising from the estimate of the heated area. The primary objective of the series of experiments reported here was to determine the fraction of O2 released into the gas phase in the 'A, state upon the thermolysis of 2PVNE. To this purpose D M F or TME was allowed to selectively react with 0, lA, and the extent of this reaction was assayed by mass spectrometry. The results of this analysis were compared when TME and DMF were used as singlet oxygen titrants in the determination of fractional 0, 'A, yield derived from 19 nm thick 2PVNE films. When TME was used, a fractional O2 lA, yjeld of only 0.1 1 was measured, whereas 0.3-0.4 was measured with DMF. The substantially lower yield found for TME is not unexpected; the rate constant for reaction of O2 IA, with TME is a factor of 20 slowers than with DMF and a slower reaction rate provides more time for the quenching of 0, 'A, by the buffer gas, by TME itself, and by the walls of the sample chamber. Even though the sample chamber was constructed of Teflon which is a very inefficient 0, IAg quencher, the surface of the endoperoxide coated plate present in the sample chamber is probably a much more effective quencher and could possibly have competed with the TME reaction for deexcitation of singlet oxygen. The decay time of the 0, 'A,-titrant reaction is equal to the inverse of the product of the density of DMF or TME (assuming that enough titrant is present so that its density remains approximately constant over the course of the reaction) and the reaction rate constant. The amount of titrant used in these exmol/cm3. Using a rate constant periments was about 2.5 X of 1.5 X cm3 mol-' s-' for DMF and 7.7 X IO8 cm3 mol-' s-' for TME, an e-folding time for singlet oxygen of 0.03 and 0.5 s is calculated for DMF and TME, respectively. Since the yield of O2 I.1, as measured with TME is substantially less than that measured with DMF, it appears that quenching in our experiments competed with the TME titration reaction limiting the lifetime of 0, )Ag to less than 0.5 s. The extent of reaction of O2 'A, with DMF was measured by using two methods. The primary oxygen mass peak at m/e 32 and the primary DMF peak at m / e 96 were used to analyze the fraction of singlet oxgyen produced by thermolysis of 2PVNE. Agreement between the two methods was not completely satisfactory; the fraction of singlet oxygen calculated from m / e 32 data is 0.4 which is substantially larger than 0.3 calculated from the m / e 96 data, though the standard deviation on the former, 0.08, is comparable to the extent of disagreement. The assumption made when calculating the fractional yield from m / e 32 data is that only O2contributes to the mass spectrometer ion current at this mass peak. It is possible that DMF fragments also contribute significantly to the ion current at m / e 32 and that consumption of DMF by singlet oxygen could result in a value for the fractional yield which is considerably larger than the true value. The primary DMF peak at m / e 96 is less subject to interference by fragmentation of other molecules. The only possible source of interference at this mass peak is the reaction product of DMF with singlet oxygen which was shown to contribute less than 5% to the
Twarowski and Dao change in ion current at m l e 96. The fraction of oxygen entering the gas phase in the lA, state is measured by the decrease in ion current signal at the primary DMF mass peak depends markedly on the thickness of the 2PVNE film. A simple model for O2 'A, diffusion and quenching in the polymer film is found to agree very well with the mass spectrometry results. The model employs a fitting parameter, k,/D, which is the ratio of the pseudo-first-order rate constant for singlet oxygen quenching by the polymer to the diffusivity of singlet oxygen in the polymer. The value of this ratio when a good fit is achieved with the experimental data is determined by the values which are assumed for two other parameters, the intrinsic yield of 0, IAg, 4, and the probability that a singlet oxygen molecule will be quenched when it encounters the Nesatron surfacepolymer interface, y. The 0, 'Ag fractional yield at the limit of zero film thickness is found to be close to 0.5 in the present series of experiments. This constrains $ to values between 0.5, when y = 0, and 1 when y = 1. When both 4 = 1 and y = 1, k, must be 2 X lo', cm-, for a good fit with the experimental data. From this value and an estimate of the diffusivity, the lifetime of O2 'A, in the polymer can be calculated. The diffusivity of O2 in 2PVNE has not been measured but the diffusivity of ground-state O2in polystyrene has been reported and will be assumed to provide an order of magnitude estimate of the diffusivity sought. In polystyrene at 88 "C the diffusivity of oxygen is 1 X cm2 s-I. With this value for the diffusivity of singlet oxygen in 2PVNE a lifetime ( l / k q )of 0.5 ps is calculated. An equally good fit of the model calculations with the data is found when 9 = 0.5 and y = 0; however, inspection of eq 1 shows that with those parameter values k , / D must be reduced by a factor of 4 in order to fit the experimental data. Therefore, the lifetime of O2 'Ag in the polymer film must be increased by a factor of 4 to give 2 M S . The range of values calculated for the lifetimes of O2 'Ag in 2PVNE solid (0.5-2 ps) is about an order of magnitude less than the 0, 'A, lifetime measured in hydrocarbon solvents6 (20-30 ps). Because the poly(vinylnaphtha1ene) matrix is expected to provide a solvent environment similar to hydrocarbons such as benzene, the faster quenching of 0, 'Ag in the 2PVNE film may be the result of a large quenching rate constant associated with the endoperoxide moiety. However, it will be remembered that the actual lifetime of 0, 'A, in 2PVNE has not been determined to better than an order of magnitude. A more direct measure of the quenching of 0, 'A, by endoperoxide is needed. In previous work the fractional yield of 0, 'A, released from 2PVNE films was calculated from photometric measurements of emission at 1.27 km and examined as a function of 2PVNE film thickness. Agreement of the photometric results with model calculations and with the DMF titration results is quite good. The same values of k , / D used to fit the titration data also furnish a good fit between model calculations and photometric results. The extent of agreement is somewhat surprising because the two sets of experimental results were not only determined by different methods but the data were collected at different thermolysis temperatures, 88 "C in the case of the DMF titration results and 130 "C in the case of the photometric results. However, the fractional yield of 0, 'A, is not expected to depend strongly on temperature if k , / D does not, and because k,/D is a ratio of two quantities which are expected to increase weakly with temperature, the fractional yield may be relatively constant over the 42 "C difference in temperature between the two sets of results. From the agreement of model calculations with fractional 0, 'A, yield results it is not possible to determine the extent of O2 'Ag quenching at the polymer-Nesatron interface since a good fit is obtained at both extremes of the range of the parameter y. Overcoating the Nesatron surface with silicon monoxide was found to have no observable effect on the fraction of O2'A, released into the gas phase. From this we conclude that surface trapped or free charge carriers are not involved in the quenching of O2 (6)Monroe, B. M. In Singlet 02;Frimer, A. A,, Ed.; CRC Press: Boca Raton, FL. 1985; Vol. 1, Chapter 5 .
J . Phys. Chem. 1988, 92, 5297-5302 'A, at the Nesatron-2PVNE interface. It may be that the quenching probability of 0 2 'A, at the Nesatron surface is unusually high as a result of hydroxyl groups attached to the metal oxide surface. If this is the case a silicon monoxide overcoating would not be expected to provide much protection from quenching since this surface also has surface hyroxyl groups which could participate in the quenching of O2 'A,.
5297
Acknowledgment. This work was supported by the Air Force Weapons Laboratory, Kirtland Air Force Base, N M 87117, under Contract F29601-85-C-0027, and performed at KMS Fusion, Ann Arbor, MI 48106. We thank Lisa Good for her technical assistance. Registry No. DMF, 625-86-5; 02,7782-44-7.
Particle Densities in Radio-Frequency Discharges of Silane Kenneth G. Spears,* Rodger P. Kampf; and Timothy J. Robinsonf Department of Chemistry, Northwestern University, Evanston, Illinois 60208 (Received: December 29, I987)
We use an rf discharge to test a new method of using fluctuations in light-scattering signals to provide particle number densities without assuming particle optical properties or shapes. The plane parallel electrode geometry for an rf discharge of silane in argon provides particle sizes which are very dependent on spatial position (0.1 mm sensitivity) and which have densities ranging from lo6 to lo8 particles/cm3 for estimated radii of 16 nm at a lo8 density. The discharge conditions and spatial positions were varied to probe low particle density and high particle density limits of the theory. Under favorable conditions of low particle number in the scattering volume, surprisingly narrow particle size variations of - 5 % were inferred from the unique distributions of light-scattering intensity.
Introduction We have previously reported laser diagnostic studies of capacitively coupled rf discharges containing silane in argon.' These studies demonstrated that small particles exist in unusual distributions between the plane parallel electrodes of our discharges. Spatially narrow zones of strong light scattering were associated with the ion sheaths. The particle size and number density both contribute to the light-scattering signal, and new techniques are required to separate these contributions. We have briefly described an approach to this problem in earlier work,2 and our accompanying paper3 fully develops the theory. We now apply this theory to silane discharges in order to analyze particle number density behavior in these unusual discharges. The ultimate goal is to study the nonequilibrium spatial chemistry that leads to particle nucleation and growth in discharges. These studies will need to be done without assuming optical properties of the particles. Chemical analysis of the particles can be done by extending our laser ablation/fluorescence technique' to elements other than silicon.
set of four fixed width slits were mounted across the monochromator slit (of a 3/4-m single monochromator (Spex)) for a slit height definition of 0.1,0.5, 1.0, and 2.0 mm. The entire chamber translates under computer control so that positions between the electrodes can be scanned. The laser is a Q switched, CW Nd:YAG laser (Control laser 512QG) equipped with an intracavity second harmonic generator of 532 nm. The pulse widths are 150 ns fwhm and the laser is usually operated at 0.1-0.5 mJ/pulse and 15 Hz. The scattered light is detected by a nine stage photomultiplier (EM1 9781R) set up for peak currents of
